Pipeline parameters different behavior in comparison to v2

[d-vesely]

PySDK Version

• PySDK V2 (2.x)
• PySDK V3 (3.x)

Describe the bug
In v2, it was possible to pass pipeline variables to environment variables into an Estimator. For example, an environment variable "RANDOM_STATE" could be set to a value given as a pipeline variable. This does not work in v3 in a ModelTrainer instance, resulting in a validation error: ValidationError: 1 validation error for ModelTrainer. A similar issue exists in the HyperparameterTuner which does not accept a pipeline variable for the random_seed, returning ValidationError: 1 validation error for HyperParameterTuningJobConfig.

To reproduce
Create ModelTrainer instance and pass a pipeline parameter into the environment argument dictionary.
Create a HyperparameterTuner instance and pass a pipeline parameter into the random_seed argument.

model = ModelTrainer(
    source_code=source_code,
    compute=compute,
    networking=self.networking,
    base_job_name=base_job_name,
    training_image=self.image_uris["train"],
    output_data_config=OutputDataConfig(
        s3_output_path=Join(
            on="/",
            values=[
                self.s3_uri_runtime,
                ExecutionVariables.PIPELINE_EXECUTION_ID,
                "02_mt_output",
            ],
        ),
        kms_key_id=self.aws_params["kms_key_hub"],
    ),
    stopping_condition=StoppingCondition(max_runtime_in_seconds=28800),
    role=self.aws_params["exec_role"],
    sagemaker_session=self.pipeline_session,
    environment={
        "RANDOM_STATE": self.pipeline_params["RandomState"].to_string(),   # <-- This line causes issues
        **self.default_env_vars,
    },
)

hyperparameter_tuner = HyperparameterTuner(
    model_trainer=model,
    base_tuning_job_name=base_job_name,
    metric_definitions=metric_definitions,
    objective_metric_name=self.hpt_params["objective_metric_name"],
    objective_type=self.hpt_params["objective_type"],
    hyperparameter_ranges=self.hpt_params["hyperparameter_ranges"],
    max_jobs=self.hpt_params["max_jobs"],
    strategy="Bayesian",
    max_parallel_jobs=4,
    random_seed=self.pipeline_params["RandomState"], # <-- This line causes issues
    tags=self.tags,
)

Expected behavior
Pipeline variables should be able to affect environment variables, as well as the random_seed argument of the HyperparameterTuner.

System information
A description of your system. Please provide:

• SageMaker Python SDK version: 3.5.0

Additional context
This is a roadblock for us regarding a migration from v2 to v3.